CoreWeave is a Generative-AI cloud platform on the frontlines of the AI infrastructure megatrend. As a new IPO (March 28, 2025), the company basically helps hyperscalers get up and running fast with early access to leading Nvidia GPUs optimized for AI. Microsoft is their biggest customer at 62% of revenues, but OpenAI signed a big deal in March that should make the future revenue split about 45% OpenAI and 23% Microsoft. This report reviews the business model, market position, growth prospects, valuation and risks (including tariffs and macroeconomics, to name two), and then concludes with my opinion on investing.
About the Business
Founded in 2017 in Livingston, New Jersey, by Michael Intrator, Brian Venturo, and Brannin McBee, CoreWeave began as a cryptocurrency mining operation under the name Atlantic Crypto. It pivoted to cloud infrastructure in 2019, leveraging its expertise in GPU management to cater to AI and machine learning workloads. Today, CoreWeave operates 32 data centers with over 250,000 GPUs, providing high-performance computing resources to enterprises like Microsoft, OpenAI, and others.
Unlike traditional cloud providers such as Amazon Web Services (AWS) or Microsoft Azure, CoreWeave specializes in GPU-accelerated compute, claiming speeds up to 35 times faster and costs 80% lower for specific tasks. Its infrastructure supports generative AI, large language models (LLMs), and other compute-intensive applications, making it a critical enabler of the AI revolution.
Worth mentioning, Nvidia backs CoreWeave through both investment and strategic partnership. Specifically, Nvidia has invested over $100 million in CoreWeave, and held a 6% stake in the company as of 2024.
Additionally, Nvidia provides CoreWeave with priority access to its GPUs, including the H100, H200, GH200, and Blackwell chips, which are critical for CoreWeave’s AI and high-performance computing services. This access gives CoreWeave a competitive edge over larger cloud providers like AWS, Microsoft, and Google, who are developing their own AI chips. Nvidia’s support also extends to engineering and infrastructure assistance, notably in projects like the $1.6 billion supercomputer data center in Plano, Texas, described by Nvidia as the fastest AI supercomputer globally.
Market Share
CoreWeave’s market share in the broader cloud computing industry remains small compared to giants like AWS, Azure, and Google Cloud, which collectively dominate with over 60% of the $480 billion market. However, in the niche GPU cloud segment—particularly for AI workloads—CoreWeave has carved out a significant presence. Sacra estimates its 2024 revenue at $1.9 billion, a 730% increase from $229 million in 2023, suggesting it’s capturing a growing slice of the AI compute market. With major clients like Microsoft (62% of 2024 revenue) and OpenAI ($11.55 billion in future contracts), CoreWeave’s traction among AI leaders bolsters its competitive edge. Still, its market share is dwarfed by hyperscalers, positioning it as a specialized contender rather than a broad-market leader.
Total Addressable Market
CoreWeave operates within the rapidly expanding AI compute infrastructure market, which Bloomberg Intelligence projects will grow from $79 billion in 2023 to $399 billion by 2028, a compound annual growth rate (CAGR) of 38%. This includes $330 billion in training infrastructure, $49 billion in inference infrastructure, and $20 billion in workload monitoring—all areas CoreWeave targets.
The broader cloud computing market, valued at $480 billion in 2024, provides additional upside, though CoreWeave’s focus remains on GPU-centric AI workloads. As AI adoption accelerates across industries like healthcare, gaming, and automotive, the total addressable market (TAM) offers substantial room for growth, potentially exceeding $500 billion by the decade’s end if current trends persist.
Growth (AI Megatrend)
CoreWeave’s rapid rise is tied to the AI megatrend, which has transformed industries since ChatGPT’s debut in 2022. CoreWeave’s revenue surged 737% to $1.9 billion in 2024, outpacing even Nvidia’s peak growth quarters, driven by demand for GPU compute from cloud providers, LLM developers, and AI-integrated apps.
Projections suggest $8 billion in revenue by 2025, fueled by a $17 billion contract backlog, including a multi-year deal with Microsoft and $11.55 billion from OpenAI. The AI megatrend shows no signs of slowing, with global AI spending expected to double by 2027.
CoreWeave’s early mover advantage—delivering Nvidia’s cutting-edge H100, H200, and GB200 clusters—positions it as a key infrastructure provider. However, its growth hinges on sustained AI investment, which some analysts question amid concerns of an “AI bubble.”
Valuation
CoreWeave IPO’d (initial public offering) at a $23 billion fully diluted valuation, raising $1.5 billion at $40 per share—below its initial $35 billion target. As of April 7, 2025, its stock climbed to $52.57, reflecting a market cap of nearly $25 billion.
Trading at a price-to-sales (P/S) ratio of less than 7 based on 2024 revenue, it appears undervalued compared to tech peers like Snowflake (P/S ~10) or Nvidia (P/S ~35). A forward P/S of 3 based on $8 billion projected 2025 revenue suggests further upside if growth holds.
However, its $863 million net loss in 2024, driven by $7 billion in debt and $941 million in interest expenses, tempers enthusiasm. Microeconomically, CoreWeave benefits from high demand elasticity for AI compute, but its capital-intensive model and reliance on Nvidia GPUs expose it to supply chain cost pressures.
Risks
CoreWeave faces significant risks.
Customer concentration is a glaring issue—62% of revenue comes from Microsoft, and a slowdown in Microsoft’s AI spending (e.g., not exercising a $12 billion option) could dent growth.
Its $13 billion debt load, used to fund data center expansion, strains profitability, with interest expenses outpacing operating income ($324.4 million in 2024).
Tariffs pose another threat: President Trump’s 2025 tariff policies could raise Nvidia GPU costs, squeezing CoreWeave’s margins as a pass-through intermediary.
Competition from hyperscalers like AWS, developing their own AI chips, and cheaper alternatives like China’s DeepSeek, further challenge its moat.
Market volatility, exemplified by a 2.7% Nasdaq drop on March 28, 2025, adds uncertainty
Capital-intensive nature of its business in a rising-rate environment is also a potential big risk.
Insider lockup period expires in September 2025, whereby insiders can sell shares (this is typical of new IPOs). If they sell it can put pressure on the shares, but if they hold it can indicated strong optimism for the shares.
The Bottom Line
CoreWeave offers a high-risk, high-reward proposition. Its alignment with the AI megatrend, explosive growth, and reasonable valuation make it an attractive buy for long-term investors betting on AI’s continued rise.
However, investors should also consider substantial risks, including high market volatility, customer concentration, debt burden, tariff-related cost increases, and competitive pressures (all warrant caution). Not to mention, microeconomically, CoreWeave’s pricing power may erode if GPU supply normalizes or demand softens.
If you are an aggressive mid to long-term investor (3-5 years+), with a high tolerance for volatility and uncertainty, the shares are compelling at current levels. But conservative investors should wait for clearer profitability signals and tariff clarity.
I have no position currently, but am watching closely and may initiate a small position on any significant price weakness.
At the end of the day, you need to do what is right for you.